Tobacco Documents Online

BAT Co LTD - MINNESOTA TOBACCO LITIGATION

P~In0RT NO. : ~ D. ~ |-7 CATEGORY: PROJECT JO~ NO.-5~.05.O1 FILE NO.: l AUTHOR: RAC~ ~C~ GROUP LEADER: )[ c~ Check Lilt Cover: i -~'~ Pho~oKraphs No. : ReSuotions : ISSUER: ~ ~v~ FS (~) : AJK (I) TAPE REy: : RAS (2) : RMC (I) : RSI; (3) : :.h~Er., F.. & D.D. Aust. : CJP de S (i) : DGF (I) : Library (2) : F£1e (2) : I]S (I) (1)! (4) Pre~e mash:or: CozTec¢ ion~ (i) Pzepa=e 1 ~our~ Co~" i7 Da~e Date I Date Sent RecdJ Reded Checking Author I~te Issuer ~te This p~ ~ is in, erred to £acilit~te and deta/1 the ;:~reso ot G.R. & D.C. B~-ts ~z~ the tTpes~-~ipt s~a~e until the ~eport is ~inally s~ved ~ issued. The I~o f~ should be co~menced at ~he time o£ 1~he i~la.l ~I=~ and &ccc.mpa~ l~he %~r----i1~ (and =as~e~ a~e~ ~=.epa.-ati~n) a~ all rabsequen~ s~ages. I~ ~ust be se~ ~o eesr~-al £tl:L~g vl~r~ the m~T wh~ ~he report is issued. ~ot_...~e: I. %~e.Tever possible, ~ts £o~ %'ypln~ ~hould be ~rri~ten in ink. 2. After %'ypi~4~, the %~j~pes~ipt should normally be re.u---ned %o 1;he AuT/~:~, who will ~ubsequen~ly Obt:aiu ~he G~p Le~e~T'I and Imm~'I &~v~l. CLTculation should be determined by ~he Issuer. 6. The t-ype$crip~ ~us~ no% be 1~Tesented £o~ p=intin~ ~il the Zesu~r has inA~ialle~ ~he p=~ £omma a~ "P~epa~e ~bu~ter". ~. A~e~ p~in~c:Ln~, one bound co1~y ~AS~ be returned to the Imsuer for final almpz~va! and the 7ssuer will ensure tha~ one copy is seen by ~he Mana~rr. Cr~ BAT Co LTD - MINNESOTA TOBACCO LITIGATION

AT'fENTIOR Thla work was pro,lucud on the H~zgnctic Ca~d typcwr[l~cz. Ir is imperativa that all correcticns, alterations, ~tc., arc made clearly. E.G. ~o,,cil}cd co,:r,.a_~s, full-stops could be easily ndssed. A cro$s in =he margin, 1,, line with the co~'rccLion would draw our aLtontion re th,2 insertion. i~ottid yOU ple.,bc l,c hind c,ou[~.h to wz'~Lc" ]~,r~,...~e ill.~orLi¢,ns ¢,n o IL~i, nr, L~. ~hcct of p.~.t,t'r, and n__o~, cut: oui:: porLionr oI~ typed worl:. ?'his J.c l~¢.c;lusc tl,u tls~l,crJc Car..ls di£fu." from ]'ap.::t- Tapes. BAT Co LTD - MINNESOTA TOBACCO LITIGATION -,..j

INDEX HAS INDICATED GAP IN BATES RANGE HERE

PROJECT COWPER - EXAMm&TION OF EXPERIMENTAL BLENDS REPORT NO. RD.1417 RESTRICTED 1.10.1976 AUTHORS: R.A. Crellin G. 0. Brooks ISSUED BY: D.J. Wood PROJECT JOB NO: 33.03.02 DISTRIBUTION: ii Dr. S.J. Green Copy No. 1-8 Dr. D.G. Felton " " 9 Library " " 10, 11 COPY NO: BAT Co LTD - MINNESOTA TOBACCO LITIGATION ¢..m .r,,,.a

INDEX HAS INDICATED GAP IN BATES RANGE HERE

T~C l GOB IT~I I Group Research & Development Centre, British-American Tobacco Co. Ltd., SOUTHAMPTON. 1st October 1976 PROJECT COWPER - EXAMn~ATION OF EXPERIMENTAL BLENDS (Report No. ltD.1417 Restricted) v SUMMARY Exper/mental flue-cured and blended ciKarettes provided by U.T.C. (South Africa) have been evaluated by normal blend and smoke analyses, O/WP analysis and taste and flavour subjective assessment. U.T.C. require Co modify EMBASSY/B. & R. in such a way as to match Rothmans smoke attributes (flue-cured experiment) whilst LUCKY STRIKE modifications should result in Lexinston smoke attributes (blended experiment). The results of the analyses described in this report indicters thac the flue-cured blend and smoke chemistry objectives could be reached by increasing the domestic flue-cured content of the blend at the expense of imported flue-cured and domestic liKht air-cured tobaccos. The PCL which was examined appeared flue-cured in char&ater and did not alter the O/WP pyridines/furfurale balance. Full O/WP analysis suEEested some increase in irritation might result from i~s usa, however. The blended objectives (blend and smoke chemistry) could be reached by increasin8 the content of domestic and/or imported Burley in the blend. Adjustment of the remainder of the blend (less domestic f~ue-cured BAT Co LTD - MINNESOTA TOBACCO LITIGATION r~

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-2- and more domestic llghC air-cured) may have assisted in achievlns the objective. In this case, since the added PCL in these samples was flue-cured in type, th£s could not be incorporated without destroying the match. PCL prepared from blended operations mighc, however, be accepuable. Long term panel tests are now in progress in South Africa to evaluate these proposals. V BAT Co LTD - MINNESOTA TOBACCO LITIGATION r,o

INDEX HAS INDICATED GAP IN BATES RANGE HERE

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-3- INTEODUCT TON An exam£naC£on (i) of the South African market revealed a distinct d£ffe~ence between brands produced by United Tobacco Co. (South) (a B.A.T company) and brands produced by Rmabrandt Group Ltd. (RoChmans Znternation81 Group). The investiKation included all brands which had a lZ or more share of the market. The resulting selection consisted of both "bleuded" and "flue-cured" types from both companies in a total of 21 brands. The brands were examined by: (a) (b) Co) Oil/wa~er partition analysis. Standard brand analysis. Smoke panel assessment of Lmpact and irri~tlon. Ze was possible to Kroup the brands by manufacturer and blend type usinK the experimental data. The Rembrandt brands were shown to be of ewo dlst£nct types: (i) "blended" produces characcarised by high O/~'P pyridines, extraceable n£cotine snd impact coupled with low O/~TP furfurals and =educing sugars and (i£) "flue-cured" products character£sed by moderate O/WE pyridines, low extractable nicoeine and impact coupled with hiKh O/WP £urfurals and reducing sugars. The U.T.C. products generally fell between (i) and (i£) in produce terms; furthermore, no clear analytical distinction be~weem U.T.C. "blended" and "flue-cured" products was found. ~ollowinK ~his clarlfica~ion of the product differences between manufacturers, experimental products were received from Soutb Africa for evaluanLon. The objective of this experLment was to modify the BENSON & HEDGESfEMBASSY blend so that one or both of these brands could BAT Co LTD - MINNESOTA TOBACCO LITIGATION k.,m

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-4- be matched to Re=he•us KS in product ceres; additionally the LUCKY STKXKE/HUNTER blend was also modified in an attempt to achieve LexinEton LS product characteristics. The experfmental products were examined by the same cechniques, (a)-(c) above, which discriminated between manufacturers and blend types. In addition, some smoke vapour phase analyses were conducted. As well •s the above experiments involvinE blend revisions, some further experiments were conducted to explore the effects of chanEes in cellulose acetate tow. EXPEEIMENTAL SAMPLES The blend compositions of the experimental ciEarettes which have been examined are listed in Table I. In addition, the composition of EMBASSY (and BENSON & HEDGES) and LUCKY STRIKE (and HUNTER) at the outset of the project, are shown in she Sable. During the course of the project, two fur=her s•ts of co~ercial EMBASSY or B. & H. and LUCKY STRIKE were received, which had incorporated some of the modifications suaEested by the experimental tiE•retie evalu•nlons. In this report, the original commercial samples - reported in (i) - have been coded ~, whilst subsequent co---erci•l samples have been coded ~ and ~, in chronological order, reflectinK the development of the brands. The experimental smnples were 818o received in 1~t,J'o separate sets, as detailed below. First Series Flue-cured Cigarettes (EO~ Elt E2) This enabl•d, throulh E0, an examination of • major blend revision; namely the elimination of PCL and domestic light air-cured and the BAT Co LTD - MINNESOTA TOBACCO LITIGATION m O ¢..,r'1 "',,.d

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-5- replacement of a subscantlal portion of imported flue-cured by domestic flue-cured. In addition the attributes of domestic and imported flue-cured could be ascertained via examination of E1 and E2 respectively. Blended Cigarettes (L~ N.P B, R, T) This enabled through L and N (L is flavoured N), an examination of blend revisions also; in this case, elimination of PCL and incorporation of domestic light air-cured toEether with increased levels of domestic Burley. Evaluation of B, I% and T enabled the attributes of do~-estlc Burley, imported Burley and domestic liEht elf-cured respectively, to be evaluated separately. At this sCaEe, commercial brands, coded ~ in this report, were also received, which had been slightly modified alone the lines of EO and L/N. Relevant competition brands - Rottnnans and Lexington - were also re-examined. Second Series Flue-cured Cilarettes (E3p E4~ ESp E6) E6 was a re-zwke of EO. E3 and E4 tested proEressive decreases in imported flue-cured content with respect to EO/E6, whilst E5 tested re-imtroduction of PCL into E3. Blended Ci~arettes~ (NL! B2r B3r B4r BS) NL was a remake of N and L. B2 and B3 tested the effects of sliEht increases in imported flue-cured and domestic Burley, respectively, at nhe expense of the domestic liEhr air-cured. B4 and B5 were B2 and B3 with PCL add-on, respectively. At this stage, commercial brands, coded ~ in this reporn, were also received, whLch had been further modified in the EO/E6 and N/L/NL direcUion. Relevant competition brands were aEain also received. BAT Co LTD - MINNESOTA TOBACCO LITIGATION "%.0

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-6- Ln addition, filter rods of 5Y/50 and 8Y/35 filament/total denier were examined to test the effect of filter modification. The majority of experimental products were tipped with 5Y/38 fil~ars. BLEND AND SMOKE ANALYSES Physical aspects of the experimennal and branded CiKarettes are detailed in Table 2; blend and smoke information is presented in Table 3. As in the recent report (I), extractable nicotine and reducing 8uKar data has been plotted, since this plot usefully dlscri~natad between manufacturers and blend types. Consequently, this plot will help to indicate whether the experimental blends are acquiring Rembrandt-llke characteristics, and which tobacco types should be included to improve the 'match' to Rembrandt products. The following observations were made from this data'- Pressure Drop Imported (E2) flue-cured tobacco pEoducad a hiKher pressure drop tobacco rod than domestic (El) flue-cured tobacco althouKh both rods have very similar packing densities. EO i8 as expected from its 50:50 EI-'E2 composition, intermediate in pressure drop. The tobacco rod pressure drops of EMBASSY and Roth~ans resemble E2 and El respectively. This may suKgest increased domestic usaKe by Rembrandt (with respect to U.T.C.) or be due to the lower density of the Rothmans tobacco rod. Imported Burley (E) results in a markedly hisher tobacco rod pressure drop than domestic Burley (B) and liKht air-outed (T) tobaccos; this, however, m£sht be explained by ~he higher density of the R tobacco rod. BAT Co LTD - MINNESOTA TOBACCO LITIGATION r,,o

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-7- 2. Nicotine/Susars E1 (domestic) tobacco contained less n~cotine but more sugars than E2 (imported). The same directional difference was noted (I) for Rembrandt and U.T.C. flue-cured brands, the Rembrandt brands containing less blend nicotine b=t hisher sugars than U.T.C. brands. This reinforces the impression that Rembrandt make greater use of domestic flue-cured grades. The E3-E5 series was higher in sugar content than EMBASSY and E2, and well matched to Rothmans. Imported Burley contained more nicotine than domestic Burley. The Burleys and air-cured samples generally contain more nicotine than the flue-cured samples. The low sugar content of these samples is reflected in the lowered sugar content of L/N/NL, and B2-B5 which contain more of these types of tobacco than LUCKY STRIKE ~. B3 approached Lexington in terms of suKar content; the sugar levels imply--,oh greater usaKe of Burley and a~r-cured in Lexington with respect to LUCK~ STRIKE. 3. Filter Characteristics The flue-cured and blended series were examined using 5Y/38 denier acetate filters w~th the exception of gO, El, E2, which were supplied with lower pressure drop 8Y/35 filters. At a later stake of the project 5Y[50 and 8Y/35 filters were supplied for comparison with 5Y/38 filters. The variations ~ filter pressure drop and plasticiser content are shown in Table 2 and are co~manted upon later in this report. 4. TPM and Nicotine Deliveries The gO, El, E2 series show h~Kher ~ and nicotine deliveries due to the lower efficiency 8Y/35 filters on these samples. However, domes=it flue-cured (El) delivers more TPM than imported flue-cured (E2) BAT Co LTD - MINNESOTA TOBACCO LITIGATION c..r'l

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-8- and, although nicocine deliveries are well-marched, a higher proporClon of the nicotine is extractable in the case of imported flue-cured tobacco. The proportion from E1 domestic tobacco (1AZ) res~nbles ~embrandt flue-cured brands (17X) (I) whereas the imported E2 proportion (21Z) resembles U.T.C. flue-cured brands (20X) (1). E3 and E4 are reasonably well-~atched co Rothmans ~ in these respects. Eochmans ~ has increased in extractable nicotine delivery. The increased level of Burley and air-cured tobacco in L/N/NL and B2, B3 has increased the extractable nicoclne deliver7 from the previous LUCKY STRIKE level to a level in accord with LexinEton ~. B4 and B5, on the ocher hand are deficient with respect Co LexinKcon ~ in terms of extractable nicotine delivery; this may be due to the presence of PCL in these samples, but would only be expected to reduce extractable nicotine delivery if the PCL was predomiruanuly flue-cured. Whereas Eotb~ans has £ncreased in extractable nicotine delivery with respect to Rothmans ~, the reverse applies to Leucinscon ~ and ~, suKEestinK some recent loss in separation between rembrandt flue-cured and blended produces. 5. Extracnable Nicotine and Reducin~ SuKars (FiKure I) FiKure 1 summarises the above sections; most of the flue-cured experiments now approach Rot~ans in blend terms, whilst B2 and B3 are good matches co LexlnEcon. Note that in all cases the presence of PCL (ES, BA, BS) depresses the extractable nicotine delivery wlCh respect Co ES, B2, 53 respectively and causes in two instances an increase in reducinK sugar content. The PCL behaves in a stronKly flue-cured manner. (Nora Chat the fiKure shows the reKions occupied by the brands, but for simplicity, only the key brands ere plotted thls time). BAT Co LTD - MINNESOTA TOBACCO LITIGATION

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-9- 6. Nicotine Transfer Efficiency In the previous surve7 C1), the nicotine transfer efficiency was shown Co be consistently higher for Rembrandt flue-cured brands (27~) than for U.T.C. flue-cured brands (22Z). This again correlates ~ich tobacco origin, since E1 (domestic) showed a high transfer efficiency (27Z), very similar to Rothmans (28Z) whilst E2 (imported) showed a low transfer efficiency (2L~), similar to EMBASSY (22X). The value for EO (23Z) was intermediate. 7. V apour Phase Data Vapour phase deliveries (formaldehyde - HCHO and nitric oxide - NO) were determined on the first series of samples. The delivery of formaldehyde is well correlated with blend reducing sugars (Figure 2) provided the same filters are used throughout. The formaldehyde deliveries from EO, El, E2 are thus higher than anticipated by the relanlonship due to uhe adopt£on of lower pressure drop filters for these samples. Whereas formaldehyde delivery correlates wi~h reducing sugars and hence blend flue-cured content, nitric oxide delivery correlates with blend Burley/air-curad content. Consequently a plot of HCHO v NO should form a useful discriminator between brands in addition to the other plots already used. It is ~uaterestin$ to note that the resulting plot (¥ilure 5) impl.ies Kreater Burley/air-cured content in Lexington than LUCKY STRIKE. If EO, El, E2 formaldehyde deliveries are lowered to el~ninate the 8Y/35 filter effecn (without thanks in NO delivery, which is not filtered) E2 approaches B. & H. El may approach Rothmans, in formaldehyde delivery, but is deficient in NO delivery. BAT Co LTD - MINNESOTA TOBACCO LITIGATION ........................................................................... --"v

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-10- It is also interesting Co note chat the pattern of the brands (and experimental samples) is sLm£1ar for the two smoke method plots -vapour r phase and O/WP (see next section). OLL/WATER PAI~TITZON ANALYSTS In the survey of South African brands, the most useful separation of brands by type and manu£aeturer was achieved by plotcinS the proportion of pyrid£nes in the scan (believed to be related to blend Burley/a£r-cu=ed content) aKainsC the proportion of cyclic aldehydes in the scan (believed Co be related to blend flue-cured content). This type of mp produced a s~ilar separation co that achieved in the previous section usins HCHO and NO del£ver£as. F£Kure 4 shows the regions occupied by the brands, but, again got clarity, only key brands are plotted. Consequently, the pos£tiou o£ the exper~nental samples (Table 4) on Chls map is discussed f£rst in thls section. However, since the scan components which are plotted represent only a small portion o£ the complete O/hrP scan, the most promfsing exper£mental samples (based on the map) are then plotted as a complete scan against the target brand to prov£de a fuller analys£s. Flue-cured/F£rst Ser£es (EOr E1p E2~ El (domestic tobacco) plots much nearer to the Rembrandt Stoup than E2 (imported tobacco), which is near the U.T.C. froup. EO is intermed£ate in position and has proKrassmd approximately half-~ay grom ~ASSY ~ to Rotbnmns ~. The position of the samples on the map was disappo£ntins, however, in that they sugKest that Rothmans position cannot be achieved, even with wholly domestic tobaccos. Tt was subsequently d£scovered thaC these cigarettes had bean tipped with BAT Co LTD - MINNESOTA TOBACCO LITIGATION

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-11- 8¥/35 acetate filter instead of the anticipated 5Y/38 filters; moreover, ~he level of triacetin plas~icise= in these filters was extremely high (Table 2); in excess of iOZ in one case. It has subsequently been shown (2) that tr£acetin exerts high filtrational selectivity towards cyclic aldehydes; consequently the high trlacenin content of these filters is responsible for the unexpectedly low cyclic aldehyde levels. Flue-Cured/Second Series (E3~ E4~ ES, E6) This series were tipped with 5Y/38 filters which contained the target 7Z of triacetin (Table 2). The positions of E3-E5 (increased domestic tobacco content with respect to EO) indicate that all three experimental products now exceed Rothmans ~ in cyclic aldehydes content and are matched in pyridines content. The difference between EO and E6 (which is a blend remake of EO) is considerable and confirms a marked effect due to the filter differences. E6 (with 5Y/38 filters) is now within the rembrandt =one and Is a slightly better match to Rothnmns than E3-E5. In consequence to this, EO-E2 were re-assessed, replacinE the EO-E2 filters with E6 (5Y/38) filters, and this data has also been plotted. Based on this re-evaluation, EI/6 and E2/6 now plot either side of Rothmans ~ on the cyclic aldehydes axis, whilst EO/6 now properly resembles E6 and Ls a good match to Rothmans ~. This re-evaluation indicates EO/6 and E6 (both 5Y/38 filters) to he reasonably good matches to Rothmans ~. El/6 now delivers vet h£Kh levels of cyclic aldehydes. Very little difference was detected between E3 and ES, indicatinE tha~ the 12Z PCL that had been added to derive E5 from E3 was neutral in 0/WP pyridines and cyclic aldehydes terms. BAT Co LTD - MINNESOTA TOBACCO LITIGATION c..rl

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-12- Branded Products The most recent version of EMBASSY, coded ~, which incorporates some of the above changes, is considerably closer'to lothmans ~ than was E~ASSY ~. However, EMBASSY ~ is still deficient in cyclic aldehydes content, and progressing EMBASSY towards the gO/6 (E6) formulation should improve the match to Rothnmns ~. It is interesting to note that Rothmans itself has not malntained its original position; Rothmans ~ is higher in pyridines. Filter Effects The effect of 5Y/50 and 8Y/35 filters on E6 rods was also assessed, and compared with 5Y/38 already assessed. The adoption of 8Y/35 filters did not move E6 back to the EO position, since these 8Y/35 filters contained only ~7Z triacetin compared with the ~lO~ content of the 8Y/35 on gO. On the other hand 5Y/50 filters do contain a higher level of triacetin (Table 2) and a lower cyclic aldehydes level results. However, it can be seen that the level of cyclic aldehydes is in the order 8Y/35 > 5Y/38 • 5Y/50 i.e. increased filtration e£ficiency lowers the cyclic aldehyde level. The varlation due to changes in tow specification does not enable E6 to be moved back into the U.T.C. zone; tow specification effects are therefore second order with respect to blend effects; plasticiser affects, on the other hand, are a first order factor. Blended/First Series (L, N, B, R, T) L and N have moved significantly from the LUCKY STRIKE ~ position towards Lexington ~, in terms of O/WP pyrldi~es content. However, the Eembran~t blended group contained still higher levels of pyridines than present in L and N. The cyclic aldehydes content has remained steady; BAT Co LTD - MINNESOTA TOBACCO LITIGATION

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-13- this was required since LUCKY STRIKE ~ and Lexington ~ had the same O/~P cyclic aldehydes content. The positions of B, R, (high pyridines) and T (moderate pyridines) suggest that an i~proved match could be achieved by the inclusion of greater quantities of Burlmy tobacco (domestic or imported) but not by the inclusion of l~ghr air-cured tobacco. Blended/Second Series (B2, B3~ B4~ BS, NL) Samples B2 ~nd B3 (which have increased levels of domestic and~or imported Burley, with respect to L/N) are very similar in pyridines/cyclic aldehydes proportions to Lexington ~. Changes in top flavourings (B3A v B3B v B3C) have no effect on these O/WP parameters. However, in this series, the add-on of 12Z PCL co B2 and B3 Co derive B4 and B5 respectively, has signif£cantly reduced the pyrld~nes levels and slightly increased the furfurals levels; thus B4 and B5 are not well matched to Lexington ~. The effects of PCL on both this series and the second flue-cured series are consistent with the hypothesis that flue-cured PCL has been used throughout. NL, a remanufmcture of N, is similar co N and BS. Branded Products The revised LUCKY STRIKE, coded ~, has progressed only marginally from LUCKY STRIKE ~ to Lexington ~, further grade changes of the B2/B3 type are required to complete the shift. Lexington ~ has declined slightly in pyridines content since the origlnel examination; this substantiates the earlier conclusion that the separation beEween Rembrandt blended and flue-cured products has declined. BAT Co LTD - MINNESOTA TOBACCO LITIGATION c...r'l k.r'I rx.) t.J"l t...r'l

INDEX HAS INDICATED GAP IN BATES RANGE HERE

In addition co the pyridines/cyclic aldehydes map, some interesting complete scan comparisons have also been included:- Em ssY 0 t:l,o o, 0 EO and Rothmans E6 and Rothmans E1 and E2 E5 and E3 LUCKY STRIKE ~ and Lexington L and Lexington B3A and Lexington ~D B and E B and T E6 (8Y/35 filters) and E6 (5Y/50 filters) (Figure 5) (Figure 6) (Figure 7) (Figure 8) (Figure 9) (Figure IO) (Figure 11) (Figure 12) (Figure 13) (Figure 14) (Figure 15) In these comparisons, it is the ratio of the absolute peak areas for the two brands which is plotted (in the previous 'map' it was the proportion of each chemical group in the scan which was plotted). RaEios within ± 0.9 are within ~he bounds of experimental error; ratios of less than 0.8 begin to be maan/nEful in terms of smoke taste. A peak identification lis~ is given in the Appendix; identified peaks are shown by full lines in the figures. The following ¢ommen~s on Figures 6-16 can be made:- EMBASSY ~ and Rothmans ~ (Figure 5~ Before any blend changes were made, EMBASSY was deficient in almost all O/WP components (except some very volaEile acyclic aldehydes and ketones). The deficiency is greatest in the seal-volatile region (approximately peaks 25 to 70), especially cyclic aldehydes and pyridinea. BAT Co LTD - MINNESOTA TOBACCO LITIGATION

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-15- EO and Rothumns ~ (Figure 6) Large changes have taken place; moving EO closer to Eothmans, but still noc matchlnK the semi-volatile levels. The EO blend also tends to increase the level of vapour phase componmnts (approximately peaks 1-25) which may be associated with irritation. E6 and Rothmans ~ (Figure 7~ A situation has now bean reached in which E5 slightly exceeds Rothmans in volatile, semi-volatile and low volatility (70-103) regions i.e. the proportions of nearly ell components are well matched to l%othmsns, E1 and E2 (Figure 8> Although cycl~c aldehyde levels are higher in domestic tobacco (flue-cured), imported flue-cured delivers more of almost all ocher O/WP constituents in all volmtiliCy tenses. E5 and E3 (¥ilure 9) Although E5 (E3 ÷ FCL) and E3 are extremely well matched over most of the scan, some increases in very volatile O/WP constituents - possibly implicated with irritation - have occurred due to PCL addition. LUCKY STRIKE ~ and Lexington ~ (Fiture SO) Before any blend changes were made, a marked deficiency in pyridines and nicotine was noted for LUCK~ STRIKE. In other respects, the profiles are well matched. L and LexlnEton ~ (Figure 11) A deficiency in pyridines still remains, although this is certainly not as large as in the case of LUCKY STKIKE ~D. L is also deficient in O/WP nicotine and delivers an excess of volatile aldehydes and ketones (peaks 9-15) which may be associated with subjective irritation. V~D BAT Co LTD - MINNESOTA TOBACCO LITIGATION

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-16- B3A and Lexington ~ (Figure 12) A situation has now been reached in which B3A 81iKhcly exceeds Lexington in nhe pyridines region. Cyclic aldehydes and keCones are well matched and che level of vet•tile aldehyde and ketones is slightly greater in Lexingnon. O/WP nicotine deliveries are pr•cisely marched. Reference Co the quality of march in Figure 9 will indicate that B3A and Lexington are by no means perfecnly matched, although pare of the mismatch (peaks 84 and 102, tri•cetin and coumarin, respectively), is not blend related. B and R (Figure 13) Substantially higher levels of O/WP pyridines and niconine are der£ved from imporued Burley (R) than from the domestic Burley (B). On the ocher hand, the level of furfural (peak 52) was much higher from uhe domesulc Burley. B and T (Figure l&) Many marked differences were observed betw•en domestic Burley and light air-cured (T). Whilst the Burley sample delivered the highest level of pyrldine8, the light elf-cured sample delivered subsCanci•lly greater quanUiUie8 of volatile aldehydes and kenones (peaks 9-16), cyclic aldehydes (peaks 52 and 60), cyclic ketones (peaks 30. 31, kA, A6, 50, 58, 65 and 82) and phenols (peaks 7A, 78, 80, 89 and 90). B delivered more indole (peak 101). E6 (8Y/35 filters) and E6 (5Y/50 filters) (Figure 15) Change from a low TPM efficiency filter (8Y/35) to • high TPM efficiency filEer (5Y/50) naturally has no effect on the very volanile components of the scan (peaks 1-35) which do nee interact with this type of fileer. The low eff£ciency filter transmits more low volatility BAT Co LTD - MINNESOTA TOBACCO LITIGATION r',,o 0,%

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-17- components (peaks 80-103 approxlmately); this is again co be expected, since these materials are probably associated with TPM. The greatest enhancement using 8¥/35 is. however, in the semi-volatile region. Some effect due to triacetin cannot be discounted; 5YfSO contains a high level (Table 2) and the O/WP triaceCin peak (84) is enhanced by 5Y/50. O/WP nicotine is little affected by the filter change. Discr iminant Anal~sis The O/WP data on the flue-cured brands originally surveyed (coded ~) together with EMBASSY ~ and Rothmans ~ were subjected tO discriminant analysis along with B.A.T Group and Rothmans brands drawn mainly from the Dutch, Canadian and New Zealand markets (4). The analysis (Figure 16) revealed a clear discrimination between the flue-cured products of the two companies, in spite of the variety of markets from which the brands were drawn. It was therefore of interest to decide whether any of the U.T.C. experimental flue-cured products had acquired Rothau~ns characteristics as defined by the discriminant analysis. The value of the dlscriminant function for the experimental products is shown on the figure. It can be deduced from the position of these samples that:- (a) use of 1OOZ domestic tobacco (El) (domestic to South Africa) provides genuine RothBans characteristics, (b) increased levels of South Africa flue-cured (E3, E4, E0, E6) improve the match to Eothmans (with respect to EMBASSY ~), (c) the low pressure drop/high triacetin filter of EO appears particularly successful., BAT Co LTD - MINNESOTA TOBACCO LITIGATION -....j r~j

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-18- (d) use of IO0Z imported tobacco (E2) retains the B.A.T Group character is tics, (e) addition of PCL (ES) to E3 causes a shift towards B.A.T Group characteristics. SMOKE PANEL ASSESSMENT The exam£natlon of brands (I) showed a separation of Kembrandt products into low impact flue-cured and high impact blended. The first series of experimental samples were assessed for impact against the terser brands. The re-evaluation showed a elishtly different relationship between impact and extractable nicotine per puff (Figure 17). The EO score for impact (Table 5) is significantly lower than EMBASSY ~ but still slightly higher than Rothmans ~ which is itself sllghtly lower than the previous assessment - Rothmans (~). As in the preceding sections, E1 (domestic tobacco) is characteristic of Rembrandt flue-cured products riving a lower impact, whilst E2 (imported tobacco) exhibits U.T.C. flue-cured product character i.e. higher impact. E0 has a level of impact midway between El &nd E2. E1 tends to have slightly higher scores for irritation (Table 5) than E2, again showing the similarity between E1 and Rothmans, which had relatively high irritation scores in the previous exercise. A very limited assessment of the flavour notes of El and E2 (Table 6) indicated that E2 has higher amplitude, more flue-cured and more green character than El. The results for N and LUCKY STKZKE ~ were very close to one another (Table 5), but N has less ~npact and throat irritation than Lexington ~. The changes in impact for a given level of irritation have moved the BAT Co LTD - MINNESOTA TOBACCO LITIGATION O t,..,m

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-19- quality cowards Lexington; however the impact of N is too low to produce the smoke quality found in Lexington. SUBSEQUENT STUDIES As a result of the laboratory assessment of experimental samples described in this report, U.T.C. intend to progress F.H2ASSY ~ and LUCKY STRIKE ~ further toward their target positions - Rothmans and Lexinston ~, respectively. To this end, a flue-cured blend halfway between EMBASSY ~ and E4 was agreed for testing by the internal lone term panel. This sample had a reduced PCL content. From Fisure 5, the characteristics of this blend should be extremely close to Rothnmns ~. A blended experiment was formulated halfway between LUCKY STRIKE and B4 with a 10% PCL addition. From Figure 5, this would still be deficient in pyridines with respect to Lexington ~, but would represent a further prosression towards Lex£ngton ~ characteristics. Results to date indicate that the panel showed preference for normal B. & H. over the f~ue-cured experiment; the experiment giving 8rearer irritation. U.T.C. point out however that the test represents an attmnpt to produce a c£garette which is acceptable to both B. & H. and Rcthmans smokers, and ~hls may not be posslble. The panel's reaction to the blended exper/~ent was o£ a 18ok of 'Burley Taste' - this would be expected from Figure 5. A further increase in blend Burley content (expected to be closer to Lexington ~) will therefore be tested. REFERENCES Io 2. 3. 4. B.A.T report No. RD.II93-R, 17.3.75. B.A.T Report No. RD.1299 Restrlcced, 28.11.75. B.A.T. Report in proEress. B.A.T. Report No. RD.1403 Restricted, 23.8.76. BAT Co LTD - MINNESOTA TOBACCO LITIGATION r',o

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-20- TABLE I i BLEND COMPOSITION Sample EMBASSY (AugusU 1975) EO/E6 E1 E2 E3 E4 E5 LUCKY STEZKE (March 1975) N/L/NL B E T B2 B3 B4 B5 Domestic Flue-cured 14 40 80 0 4O 49 36 23 19 0 0 0 Air-curd 19 19 17 17 7 0 0 0 O O O 0 7 0 0 78 Zmpor~ed Flue-cured 54 43 0 85 40 31 36 Doues~i¢ Burley 5 13 78 0 0 13 16 12 14 3 O 3 0 23 21 0 O o 25 25 22 22 Impor ted Burley , J, 0 0 O 0 0 0 0 0 0 0 0 0 0 O 15 15 0 100 0 18 18 16 16 Turkish Seem PCL O 1B ~0 O17 0 0 20 0 0 15 0 O 20 0 O 20 0 0 18 lO 9 14 10 9 13 0 0 22 0 0 0 0 0 22 0 9 13 0 9 13 0 8 12 10 8 12 10 BAT Co LTD - MINNESOTA TOBACCO LITIGATION • c.J'l r',o o',.,

-21- Ssmple Fl~r-c~red $e~es ZO (SY/3S fileers) El (82/35 f£1cer8) E2 (8Y/33 f£1ters) E3 (5Y/38 £~ers) E~ (5T/36 fLlters) E~ (5Y/38 f£1cars) E6 (5Y/38 f£1tern) E6 (SY/50 f£1ter8) E6 (8Y/35 f£1=us) ~SY ~ (5Y/SO ££1ters) to r.i-maz= (D J~r.im,=, s. • u. ~ (.5Yl38 ftlce=8) I~ended Ser£es L (ST/38 f£1:ers) U (5Y/38 f£1ctrs) B (5T/38 fLLte:s) 1 (5Y/38 fLl~ars) T (5Y/38 f£1cers) Hr. (5Y/38 f£1ter8) S2 (5Y/38 f£1c*=8) S3(A-C) (SYl3s f£~ers) B4 (5Y/38 filters) B5 (5Y/38 fLlcors) ~'o~r ~ (~ (5Y/38 f~Zcer,) ~uc~ S'LT3~ ~) (5Y/3~ flUte=s) LaziuS~On 1,¢zinSeon ~i,,sr.o,, (~ TAJ~I.~ 2 lq~rs~c,u. ~jrrs T~csl FLI:e: OensL~ LanEch (l~m~h (q m-') (---) 85 20 321 8~ 20 322 84 20 325 85 2O 3O5 85 2O 319 85 20 3O6 85 20 308 85 20 308 85 20 308 85 20 318 84 20 31~ 85 18 295 8~ 18 287 J~ 18 290 85 17 304 86 17 31~ 79 15 310 I0 15 79 15 289 79 15 316 79 1S 298 79 15 302 79 25 298 79 1.5 292 ?9 15 ]07 79 15 303 79 25 )07 7~ 1~ 3O3 79 15 303 8O 1~ 199 80 15 297 79 15 392 P.Do (To r.~l ) (m V.C.) 10.3 8.6 11.1 12.2 12.3 10.7 12.7 12.7 12.7 l&.2 i.4.5 9.7 9.& 9.~ tZ.5 13.7 13.6 11.5 1&.3 19.3 10. $ 11.6 11.5 11.2 10.~ 10.5 g.a 9.6 9.9 tL1 8.9 8.7 (Tilte:) (-- v.C.) 2.? 2.3 2.3 3.9 3.8 4.0 3.8 S.6 2.5 5.6 5.6 6.1 &.2 &.l 4.7 6.3 3.8 2.9 3.3 3.$ 3.2 3.5 3.6 3.6 3.6 3.2 3.2 3.4 3.3 3.1 3.3 3.2 ]P.D. ('Z'oboc¢o) (,-- v.~.) ?.3 6;1 8.9 8.3 8.5 6.7 8.9 8.9 8.9 8.9 8.9 5.5 5.3 5.5 7°7 9.6 9.5 7.7 9.2 16.7 7.8 8.1 7.9 7.6 7 .O 7.3 6.1 6.3 6.6 6.1 5.5 S.5 Poroe £ ~:y (WZ'~) 9] 99 86 103 77 77 77 77 139 117 9~ 126 72 81 77 6~ 67 59 7O 117 112 111 107 132 79 81 123 128 169 1&8 Fi1:er T:Lacttin 9.1 10.1 9.~ 7./* 9.7 7.3 8.0 8.~ 6.~ 8.8 i i 6.4 6.8 i 8.8 8.2 8.9 BAT Co LTD - MINNESOTA TOBACCO LITIGATION (_m r'~ ........................................................................... -'---'--'v

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-22- T~I,E 3 BI.EITD AND SMOKE H~LSUKE)ENTS Samaple Flue-cured Series EO El E2 E3 E4 E5 E6 EMBASSY (~ %ASSY LI~IUD8 B. & it. ~) B. aH. ~ BLended Series Blend Nicotine (Z) 1.3 1.3 1.5 1.1 1.3 1.2 1.2 1.2 1.7 1.2 1.5 1.6 1.5 1.5 Keducins Tonal ~ Nicotine Sugars Susars (=) (Z) (mZ) (aZ) 19.2 22.5 34 1.8 19.1 22.5 42 2.1 16.8 19.2 31 2.1 17.0 19.8 30 1.4 18.0 20.0 35 1.7 17.8 19.6 36 1.6 18.6 20.5 33 1.5 14..0 17.6 25 1.& 16.0 17.4 25 1.5 17 • 0 20 . 1 34 1.8, 16.2 18.6 36 2 • 1 13.7 18.3 37 2.1 14.4 17.3 32 1.9 16.6` 18.6, 27 1.6 L N B R T NL B2 B3 (A-C) B5 1.7 9.8 1.7 10.4 1.8 3.6 2.1 3.4 1.9 4.6` 1.6 11.7 2.0 11.2 1.9 9.6` 1.8 10.3 1.8 10.4 LUCICY S'I'iLI;i~ ~) LUCET STRI~ LUCK~ STE~ Lezi.nSco- (~ Lexington (~D Lmd.nSt:on (~) 1.9 1.8 2.0 2.0 1.8 1.9 12.9 13.5 L3.6 9.0 8.0 9.3 11.5 35 2.0 11.8 30 1.9 5.0 31 1.6 4.2 35 2.3 5.8 29 1.g 13.3 37 2.1 12.2 39 2.3 10.8 36 2.1 11.6 39 2.0 11.7 39 1.8 14.9 41 2.5 15.1 39 2.6` 14.4 41 2.0 11.7 39 2.4. 11 • 3 40 2.3 10.9 4O 2.0 lxcraccable Puff N£cocine Number 0.31 0.32 0.42 O. 20 0.23 O.14 O. 17 0.32 0.27 0.26 0.30 0.43 O.42 0.34 0.55 0.53 0.51 1.03 0.44 0.34 O. 63 O. 53 0.43 0.37 0.44 0.50 0.3.5 0.58 0.68 0.37 10.7 11.5 11.4 10.6` 11,5 10.4 11.2 IO. 2 9.8 11.2 10.5 11.1 11.0 10.5 11.1 10.8 9.5 10.2 9.7 10.7 10.4. 9.9 10.0 10. $ II.0 11.0 IO. 6 11.0 10.3 10.2 Formaldehyde (~,8) 94 138 101 m o D 74 57 54 36 45 17 14 30 m 56 59 33 Nitric OXide (us) 42 72 70 98 m 68 37 146 142 226 185 172 m o 117 105 214 BAT Co LTD - MINNESOTA TOBACCO LITIGATION C~ r~o

INDEX HAS INDICATED GAP IN BATES RANGE HERE

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-23- TASLE 4 O/WP PYRIDINES AND FURFUItALS (NORMALZSED DATA) v Blend EO E1 E2 E3 E4 E5 E6 EO El E2 E6 E6 EMBASSY EMBASSY I%o ~hman• Ro l:hman• ¥il~er L N B E T NL B2 B3A B3B B3C B4 B5 LUCKY STEIE~ LUCKY STRIKE Lexington Lexington 20 mm 8Y/35 20 mm 8Y/35 20 mm 8Y/35 20 mm 5Y138 20 mm 5Y/38 20mm 5Y/38 20 mm 5Y/38 20--- 5Y/38 20 mm 5Y/38 20 mm 5Y/38 20 mm 8Y/35 20 ~n 5Y/50 20 mm 5Y150 20 am 5¥/50 D m -~ 5Y/38 ~- $¥/38 mm 5Y/38 mm 5Y/38 am 5¥/38 mm 5Y/38 mm 5Y138 ~n 5Y/38 ~n 5Y/38 I 5Y/38 mm 5Y/38 .i. 5Y/38 5Y/38 mm 5Y/38 m 15 15 15 15 15 15 15 15 15 15 15 15 15 Pyrldines (%) 8.4 8.9 9.3 9.3 9.6 9.4 9.5 9.0 9.4 9.7 9.3 10.0 8.3 9.4 9.5 11.6 13.1 12.8 15.3 16.2 13.0 12.8 14.2 14.1 14.2 14.4 13.4 12.9 11.6 12.0 14.4 13.6 Furfurals (z) 6.0 6.8 4.9 9.0 9.5 9.0 8.8 7.9 9.9 7.0 9.7 7.2 4.3 5.9 7.5 8.3 5.9 6.1 4.1 3.6 5.3 6.9 6.4 6.4 6.8 6.7 7.6 7.1 6.2 7.A 6.3 6.4 BAT Co LTD - MINNESOTA TOBACCO LITIGATION

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-24- TABLE 5 XRRITANT SENSATION S i lrri~ation ~pacc Mouth ThzoaC Nose 2.5*** 3,5 3,9 4,1 1 • 9*** 3.7 4. I 4.5 E0 2.4** 3.7 3.9 4.1 EM3ASSY ~ 3.3** 3.6 3.5 4.1 El 2.0** 3.8 4.0 4.3 E2 3,I** 3.5 3.7 4.2 N 3.5* 3.7 3.7" 4.3 Lexinscon (~) 4.2* 3.9 4.2* 4.5 N 3,3 3.6 3.7 4.~ LUCKY ST]LIKE (~ 3.2 3 . 9 3 . 8 4 . 6 r zl Difference between means significant eC ~98Z~ [95:J level of confidence. Significance of difference between mem-s assessed by Wilcoxon matched-pairs, signed-fanEs test. BAT Co LTD - MINNESOTA TOBACCO LITIGATION ¢...m r'~

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-25- TABLE 6 FLAVOUR PROF ZLE Amp i i cud • Class i Flue-cured Turkish Air-cured Blended Class 2 Greenp Grassy l~us cy Earthy D £r cy Class 3 Roasted, Toasted Burnt Add£t:ive E1 2.7 1.2 3.2 E2 3.2 1.7 3.0 Znsufficienc resulcs for scacisc£cal ana3.ysis. BAT Co LTD - MINNESOTA TOBACCO LITIGATION C..r'l

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-26- APPENDIX O/WP PEAK LDENTIFICATION LIST Peak Number 4 9 11 12 14 15 16 20 22 23 24 27 28 29 30 31 32 34 36 37 39 41 42 43 44 46 Peak Peak Marne Peak Name Number ,... ,.,, ,, ,, 71 74 75 76 77 78 80 82 84 88 89 90 98 101 Propionaldehyde Ace tone/Isobu tyTa Idehyde Me=hyle~hyl ketone Isovaleraldehyde Mechylvinyl ketone Bu~an--2,3-dione Pen tan-2-one But-2--eni tri le Bu:anitrile (Several compounds) N-Me chy Ipyrr o le Pen ta-2, ~-dien- ~-al Bu ~-3-en£ ~ri le Pyr idine Cyc lopennanone 2-Methy icyclopentanone 2-Me ~hylpyridine 2,6-Dimethylpyrid~ne 2-Methylpyrazine (tent) 1,2-Dimethylpyrrole (cent) 3- and/or 4-Me~hylpyridina 2,5-Dimeuhylpyr id£ne 2,4-Dimethylpyridlne 2,3-Dimethylpyrazlne Cyclopent-2-en-l-one 2-Met hylcyc lopent-2-en- l-one 47 50 52 53 55 56 57 58 60 62 64 55 3-E thylpyridine Cyc lo hex-2- en-l-one Furfural 3-V inylpyrid ine Hexan-2,5-~ione 3-Met hylcyclopent-2-en-l-one and Pyrrole Benzaldehyde 2,3-D ~methylcyc Lopent-2-en-l-one 5-Methylfurfural Benzonitrile Furfuryl alcohoi Acetophenone 1.2-D/methylpyrrole-5-aldehyde (cent) Guaiacol N icot ~ne 2-Phenylethanol Phenylacetonitr ~le 4-Methyl~ia¢o L Pheno ~/o-~resol Zndanone m.p-Cresol (Trlacetin) Clnnamonltr il e (tent) 3.5-Xylenol (tent) 4-~inylEuaia=ol (tent) ('t'ZGDA> Zndole BAT Co LTD - MINNESOTA TOBACCO LITIGATION OO

FIG. I RD. 14r7 RESTRIC,'r'ED EXTRACTABLE NICOTINE AND REDUCING SUGAR,5 EXTRACTABLE NICOTINE (,,,9/cIG) I.O o-G 0-4. O.P- O BAT Co SR LTD • 8Z ~REH. Bt,, SLNoT°NO U.T.C,. ~. (~) SNL / "" .... I.._. ~ El I;(o / "-"~",P" eROrHH.'4NS (D \ / oE, u~c. r/c /~"~ ~ , ~ ®E6 REM. F/C, ® ¢'5 ! I ! ; 5 IO 15 ZO REDUCING ,,~JQAR.5 (~., ¢L~rb) - MINNESOTA TOBACCO LITIGATION

FI~. 2 RD. 1417 FORHALDEHYDE AND REDUCING SUGAR8 RESTRICTED HCHo (~'9) I I 14.O 120' Io0 6o I I ! I I ! I ®E2 / /®¢o/ /J N 40 eL ~TON (~ Z0 0 4. 8 12 16 ?.,0 REDUCJN~ SUGARS C'/~ BAT Co LTD - MINNESOTA TOBACCO LITIGATION Lrl "-.M E~ ................................................................ ---.----.

VAPOUR FIG.3 PHA,SE DELIV[RIF..5 1417 RESTRICTED No (u..9) P-.50~ ZOO, 150 IO0 0 ~R = I.EXINGTON (~) ~T LUCKY O~T'RII~ Q LUCKY o ST'RIKE (~) • ROTHHANS ® oBa, H(:D''--'--O ''~ ~ oeJ~H ®..-----o¢ o | ! 5O I00 HCHo 0,~) BAT Co LTD - MINNESOTA TOBACCO LITIGATION v 150 -,.,,j Cc

% I:WRI DINE.S 1"7 OR 16 15 14- 13 12 II IO 9 8 7 3 BAT Co Fi G.4 RD 1417 PYRI DINES/CY'C, LIC ALDF_.HYDE5 MAP RE,gTP.I CTE D oB LEXIN~TONo~)D4 OT 0~ E4 . I I I I I I I I 4. 5 G "7 8 ~ IO II o~ CYCUC ALDEHYDF-~ LTD - MINNF_,SOTA TOBACCO LITIGATION

m m ~ND m~ F__-S_. m m ii ,,.# m mmmmm~ W o ~ 6 o 6 6 o~ c~ BAT Co LTD - MINNESOTA TOBACCO LITIGATION

G or) Z O i Ii BAT Co LTD - MINNESOTA TOBACCO LITIGATION

NOI£VDIllq ODDV~IOI V£O~NNII~I - Cllq oD .LV~ ii Zo C CID w m~ m ~n,n~ i , . IN • ~ o m m I"11 1"11

m I-- o~ w D w i la. oo Z m W ! A D La6 BAT Co LTD i 2 u "T ill! !! llli o~ ql,,n,J Io o'7 I0,J o,. II iI I D~ I[-.-- ~Imm i ~m m m ~m ! I 0 G 6 6 ,:,,,. MINNESOTA TOBACCO LITIGATION i I ~ anita am. ! ! ! 0 _u",,,,.

B m m c¢ la. w Z u ma i e~ mi mi m~m oa m n mnnuu n u i ~ 4iuuR, l,,o.I nn m J m 2 BAT Co'LTD - MINNESOTA TOB'ACCO LITIGATION w,Q

cO > -4 ('1 o t" -4 I Z Z > .-] O DO > C1 C1 O F" i,,,,-* -4 C~ > O z LUCKY' STRIKE > I.£XlNqTON 0"2 0.4 0.6 ii II I I I.O 10 0.4 LUCKY STRIKE < LEXINGTON C. LUCKY rm. o ( STRIKE (D AND LEXlNGTONO I I I I I I I I I I I I I I I R0.141"/RESTRICTED II0 PEAK NUMI~ER

0 W m m E Z ~m ~m . z P_. mm m m J ~a w N~mia m m~ ~mml n~moa ~o m _J Z "-'-'--- (_jr'l r~ .~ z A BAT Co LTD - MINNESOTA TOBACCO LITIGATION r,,o

~mu m I m mmmi m • i ~mmm m t.7 .J ~ ........ : ~. I ,immimmm" ,im m " | 0 ! ~f t~r) .n Z (..n Z m BAT Co LTD - MINNESOTA TOBACCO LITIGATION

°il a w u iii m um .II II iIIIIi.Imm III m IIiI | .minim. m imlI r IN m~ a ImI. iroN ....... I i, I I m I IiIii i I I ~Im ~ mm iiimIm. mw m-- ---- D IIiIIIII Unl IiI I I II. I In | m I mo~ ~m i gm mm m , a m mw D m m~ m m N Dm~ Q m~ m ! I ...... O r,,%~, o BAT Co LTD - MINNESOTA TOBACCO LITIGATION

°i "V L. ! t ' = ! A N qd~ ~ m Ill I m mm~m~ m immm,~m.i i I ! I m.m~ II m I II IIII II I I I ~ II II ~IIIII I I I II ~m I,III I I I Ill I I IIIIIII I I II I IIIIII I • o BAT Co LTD - MINNESOTA TOBACCO LITIGATION | Q , c ,rl U~

m L~ L~J <: LLJ ~ ~m~mmm ~m~w ~m eDI u m mew m m~im~ m m ~ gm,m Gm. r m ~J7 r~ BAT Co LTD - MINNESOTA TOBACCO LITIGATION o-,,

( ( RDI4.17 RESTRIOTED FIG. 16 > (--I o ! Z Z > 0 > N © t-" -4 > -4 ©. Z No. OF BRANOS -----t----I- El EO DISCRIMINANT ANALYSIS I 2 1'_,2 I z l ROTHHANS (~) AND ~) I----I--4-----t----t E4 £6 E3 E2 E5 u.t.c (sour. Arrack) EXPERIMENTAL PRODUCI5 re~ot~oumarR~:~ BRANDS UT.C.(8OUTH AFRICA) BRANOS ~THM~S (4 M~K{n) BRANOS Z6 /g}TcjO I t I 4. 5 f--'l-'l _ _ I I __ I "i I I I ! , I I 0 -I - 2 "3 "4. DISCRIHINANT FUNOTION ~.A.T. (4 MARKETS) BRAND5 I -5

IMPACT AND ,i i im FIG.17 RD. 141"I EXTRACTABLE NICOTINE RESTRJG'rED IMPACT 4..5. eGUNSTON IPAUL ® DOLLAR oB&H OF.2 EXPERIMENTAL,5 • £o BAT Co LTD - MINNESOTA TOBACCO LITIGATION m P,o cE:)

INDEX HAS INDICATED GAP IN BATES RANGE HERE

O0~lq>~qOl